Food-i-Sense Analytics: Integrating AI Into Continuous Glucose Monitoring Data Analysis for Precision Nutrition.

NCT07626658 | Not Yet Recruiting FDA Device

• 1 other identifier: IMD PI-078
• Study Type: observational
• Target: 471
• Locations: 0 countries
• Sites: N/A

Brief Summary

This study aims to improve how we understand and manage blood sugar responses in adults without diabetes. Even in people who appear healthy, blood sugar levels after meals can behave in different ways. These patterns may help predict future risk of diseases such as type 2 diabetes or other cardiometabolic problems. To study this, researchers at IMDEA Nutrition have developed a computer algorithm called GLIA, which uses artificial intelligence (AI) to analyze continuous glucose monitoring (CGM) data. The goal is to classify people into different "glucotypes", meaning typical patterns of how their blood sugar behaves throughout the day. These glucotypes could help tailor dietary recommendations in the future. Goals of the study

• Do not\*have diagnosed diabetes or serious metabolic disease.
• Agree to wear a glucose sensor for 14 days.
• Can keep stable eating habits and record diet and physical activity. What participation involves The study lasts 3 weeks and includes 3 visits: Visit 1 - Screening (20 min):
• Review of eligibility criteria.
• Explanation of the study.
• Signing informed consent.
• Visit 2 - Initial assessment (45 min)
• Collection of personal and health information.
• Measurements: weight, height, waist, body composition, blood pressure.
• Placement of a FreeStyle Libre 3 CGM sensor.
• Instructions for:
• Completing two 3-day food records (one each week).
• Taking photos of all meals.
• Reporting physical activity. Continuous monitoring (14 days) Visit 3 - Final evaluation (45 min)
• Review of diet records.
• Repeat measurements.
• Blood and urine samples are collected for metabolic and molecular analyses. Meal photos are analyzed using an AI-based food recognition model. The system identifies foods and estimates nutrients (macronutrients, vitamins, minerals, glycemic index, etc.). This helps researchers understand how meals relate to blood sugar patterns. Potential benefits: Although participants may not receive direct health benefits, the study will:
• Improve understanding of how healthy people process glucose.
• Help identify early risk markers for metabolic diseases.
• Contribute to developing \*\*personalized nutrition tools\*\* based on individual glucose responses. Risks: are minimal and mainly include:
• Mild skin irritation from the CGM sensor.
• Temporary discomfort from blood draw.

Conditions

Artificial Intelligence Mobile Application; Prediabetes (Insulin Resistance, Impaired Glucose Tolerance)

Keywords

Continuous glucose monitoring (CGM); Glucose dynamics; Glucose phenotyping; Glucotypes; Artificial intelligence; Nutrition; Machine-learning; Glucose patterns; Precision nutrition; Chrononutrition; Glycemic variability; Nutritional pattern; Glycemic response modeling; Personalized dietary recommendations; Multimodal metabolic phenotyping; Wearable glucose sensors; AI-driven health monitoring; Cardiometabolic health; Adults Without Diabetes

Outcome Measures

Primary Outcomes (1)

• Glucotype Classification Derived From Continuous Glucose Monitoring Data

  The primary outcome is the glucotype assigned to each participant based on analysis of the 14-day continuous glucose monitoring (CGM) trace. Glucotypes reflect individualized patterns of glucose dynamics, capturing peak shape, amplitude, recovery, variability, and chrononutrition-related fluctuations. The classification is generated using the GLIA machine-learning algorithm, which incorporates preprocessing (normalization, artifact detection), multivariate feature extraction (including peak morphology descriptors, temporal patterns, and variability metrics), and unsupervised clustering with stability assessment. The outcome quantifies each participant's predominant glucose-response phenotype under free-living conditions and serves as the foundation for assessing associations with dietary intake, metabolic traits, and predictive modeling of glycemic responses.

  Assessed continuously over 14 days of CGM wear, with glucotype classification calculated after completion of the full 14-day glucose-monitoring period for each participant.

Secondary Outcomes (13)

• Body mass index

  Measured during Visit 2 and Visit 3 across the 14-day monitoring period.

• Waist circunference

  Assessed during Visit 2 and Visit 3 within the 14-day monitoring period.

• Body Fat Percentage

  Measured during Visit 2 and Visit 3 across the 14-day monitoring period.

• Muscle mass

  Measured during Visit 2 and Visit 3 within the 14-day CGM period.

• Visceral Fat Index

  Measured during Visit 2 and Visit 3 over the 14-day monitoring period.

Other Outcomes (6)

• Energy Intake

  Assessed across the 14-day CGM monitoring period, combining two 3-day diet records and all photographed meals.

• Macronutrient Distribution

  Assessed across the 14-day CGM monitoring period, using two 3-day diet records plus continuous meal-photo submissions.

• Micronutrient Intake

  Assessed throughout the 14-day CGM period, based on both 3-day diet logs and all meal photographs.

Study Arms (1)

Food_iSense Analytics Cohort

This cohort includes adults aged 18-70 years without diagnosed diabetes who undergo continuous glucose monitoring (CGM) for 14 days using a FreeStyle Libre 3 sensor. Participants complete structured dietary records, provide meal photographs for AI-based food recognition, and answer validated nutrition and physical-activity questionnaires. Anthropometry, body composition, blood pressure, and recent clinical history are collected at study visits. At the end of monitoring, fasting blood and first-morning urine samples are obtained for biochemical and molecular analyses. No therapeutic intervention is administered; instead, the study characterizes natural glucose-response patterns ("glucotypes") under free-living conditions and evaluates how diet, lifestyle, and metabolic traits relate to glycemic dynamics to support future precision-nutrition strategies.

Device: Continuous glucose monitoring using a wearable sensor (flash interstitial glucose monitor)

Interventions

Continuous glucose monitoring using a wearable sensor (flash interstitial glucose monitor) DEVICE

The intervention consists of applying and wearing a 14-day continuous glucose monitoring (CGM) device that captures interstitial glucose every minute under free-living conditions. This wearable flash sensor is used exclusively for passive data collection; it does not provide insulin delivery, therapeutic adjustments, or real-time clinical management. What distinguishes this intervention is its integration into a multimodal data-capture system: participants simultaneously complete structured dietary records, submit standardized meal photographs for AI-based food recognition, and undergo detailed phenotyping. The CGM data are then processed through the study's proprietary GLIA algorithm to derive individualized glucose-response patterns ("glucotypes"). This combination of high-frequency glucose monitoring, dietary image analytics, and machine-learning modeling differentiates the device's use from typical clinical or self-management applications in other studies.

Food_iSense Analytics Cohort

Eligibility Criteria

• Age: 18 Years - 70 Years
• Sex: all
• Healthy Volunteers: Yes
• Age Groups: Adult (18-64), Older Adult (65+)
• Sampling Method: Non-Probability Sample

Study Population

The study population consists of community-dwelling adults aged 18-70 years, recruited from the general population through public advertisements, university campus postings, pharmacies, and primary care centers. Participants represent a broad, non-clinical community sample without diagnosed diabetes or severe metabolic disease. Recruitment is open to individuals living independently and able to maintain their usual daily routines. The population reflects a heterogeneous mix of sociodemographic backgrounds to ensure variability in dietary habits, lifestyle patterns, and glucose-response profiles.

You may qualify if:

• Adults aged 18 to 70 years.
• Willing and able to undergo 14 days of continuous glucose monitoring (CGM) using a wearable sensor.
• Able to maintain stable dietary habits during the monitoring period.
• Able and willing to complete dietary records, including two structured 3-day food logs.
• Able and willing to photograph all meals during the 14-day monitoring period following instructions provided.
• Able to keep a record of physical activity as instructed.
• No previous diagnosis of diabetes or other serious metabolic disorders.
• Sufficient commitment and availability to attend all study visits (screening, baseline evaluation, final evaluation).
• Capable of providing written informed consent.

You may not qualify if:

• Diagnosed diabetes mellitus or other serious metabolic disorders.
• History of severe gastrointestinal, cardiovascular, or other medical conditions that may interfere with stable diet or physical activity during the study.
• Pregnant or breastfeeding women.
• Inability or unwillingness to comply with continuous glucose monitoring (CGM) procedures for 14 days.
• Participants with skin conditions or allergies that prevent safe use of a CGM sensor.
• Current participation in another clinical trial that could affect study results.
• Use of medications that significantly alter glucose metabolism or interfere with CGM accuracy.
• Inability to attend all scheduled study visits or complete required records (diet logs, photos, questionnaires).
• Any condition judged by the investigators to make the participant unsuitable for the study or unable to provide informed consent.

Sponsors & Collaborators

• IMDEA Food: lead
• Abbott Laboratories (Pak) Ltd.: collaborator

Related Publications (8)

• Klonoff DC, Nguyen KT, Xu NY, Gutierrez A, Espinoza JC, Vidmar AP. Use of Continuous Glucose Monitors by People Without Diabetes: An Idea Whose Time Has Come? J Diabetes Sci Technol. 2023 Nov;17(6):1686-1697. doi: 10.1177/19322968221110830. Epub 2022 Jul 20.

  PMID: 35856435 BACKGROUND

• Hengist A, Ong JA, McNeel K, Guo J, Hall KD. Imprecision nutrition? Intraindividual variability of glucose responses to duplicate presented meals in adults without diabetes. Am J Clin Nutr. 2025 Jan;121(1):74-82. doi: 10.1016/j.ajcnut.2024.10.007. Epub 2024 Dec 2.

  PMID: 39755436 BACKGROUND

• Mao Y, Tan KXQ, Seng A, Wong P, Toh SA, Cook AR. Stratification of Patients with Diabetes Using Continuous Glucose Monitoring Profiles and Machine Learning. Health Data Sci. 2022 Apr 27;2022:9892340. doi: 10.34133/2022/9892340. eCollection 2022.

  PMID: 38487483 BACKGROUND

• Hall H, Perelman D, Breschi A, Limcaoco P, Kellogg R, McLaughlin T, Snyder M. Glucotypes reveal new patterns of glucose dysregulation. PLoS Biol. 2018 Jul 24;16(7):e2005143. doi: 10.1371/journal.pbio.2005143. eCollection 2018 Jul.

  PMID: 30040822 BACKGROUND

• Zeevi D, Korem T, Zmora N, Israeli D, Rothschild D, Weinberger A, Ben-Yacov O, Lador D, Avnit-Sagi T, Lotan-Pompan M, Suez J, Mahdi JA, Matot E, Malka G, Kosower N, Rein M, Zilberman-Schapira G, Dohnalova L, Pevsner-Fischer M, Bikovsky R, Halpern Z, Elinav E, Segal E. Personalized Nutrition by Prediction of Glycemic Responses. Cell. 2015 Nov 19;163(5):1079-1094. doi: 10.1016/j.cell.2015.11.001.

  PMID: 26590418 BACKGROUND

• van Doorn WPTM, Foreman YD, Schaper NC, Savelberg HHCM, Koster A, van der Kallen CJH, Wesselius A, Schram MT, Henry RMA, Dagnelie PC, de Galan BE, Bekers O, Stehouwer CDA, Meex SJR, Brouwers MCGJ. Machine learning-based glucose prediction with use of continuous glucose and physical activity monitoring data: The Maastricht Study. PLoS One. 2021 Jun 24;16(6):e0253125. doi: 10.1371/journal.pone.0253125. eCollection 2021.

  PMID: 34166426 BACKGROUND

• Barrea L, Verde L, Colao A, Mandarino LJ, Muscogiuri G. Medical nutrition therapy for the management of type 2 diabetes mellitus. Nat Rev Endocrinol. 2025 Dec;21(12):769-782. doi: 10.1038/s41574-025-01161-5. Epub 2025 Aug 15.

  PMID: 40817355 BACKGROUND

• Safiri S, Karamzad N, Kaufman JS, Bell AW, Nejadghaderi SA, Sullman MJM, Moradi-Lakeh M, Collins G, Kolahi AA. Prevalence, Deaths and Disability-Adjusted-Life-Years (DALYs) Due to Type 2 Diabetes and Its Attributable Risk Factors in 204 Countries and Territories, 1990-2019: Results From the Global Burden of Disease Study 2019. Front Endocrinol (Lausanne). 2022 Feb 25;13:838027. doi: 10.3389/fendo.2022.838027. eCollection 2022.

  PMID: 35282442 BACKGROUND

Biospecimen

Retention: SAMPLES WITH DNA

Stored plasma, blood cell fraction, and first-morning urine aliquots for future analyses.

MeSH Terms

Conditions

Prediabetic State; Insulin Resistance; Glucose Intolerance

Condition Hierarchy (Ancestors)

Diabetes Mellitus; Glucose Metabolism Disorders; Metabolic Diseases; Nutritional and Metabolic Diseases; Endocrine System Diseases; Hyperinsulinism; Hyperglycemia

Central Study Contacts

Lidia Daimiel Ruiz, Senior Researcher

CONTACT

+34655250563; lidia.daimiel@nutricion.imdea.org

Víctor de la O Pascual, Junior Researcher

CONTACT

+34648749288; victor.delao@nutricion.imdea.org

Study Design

• Study Type: observational
• Observational Model: COHORT
• Time Perspective: CROSS SECTIONAL
• Sponsor Type: OTHER
• Responsible Party: PRINCIPAL INVESTIGATOR
• PI Title: Principal Investigator

Study Record Dates

• First Submitted: May 27, 2026
• First Posted: June 4, 2026
• Study Start: July 1, 2026
• Primary Completion (Estimated): May 1, 2028
• Study Completion (Estimated): December 1, 2028
• Last Updated: June 4, 2026

Record last verified: 2026-06

Data Sharing

• IPD Sharing: Will share
• Shared Documents: STUDY PROTOCOL
• Time Frame: Individual participant data (IPD) and accompanying documentation will become available not before 12 months after completion of the final data analysis, anticipated to begin once all primary and secondary outcomes are fully evaluated. Data will remain accessible to qualified researchers for a minimum of 5 years following the initial release. After this period, continued availability will depend on dataset relevance, ethical approvals, and storage capacity. Access will be granted only through controlled procedures and under a signed data-sharing agreement ensuring secure use and strict protection against re-identification.